Copyright UNU/WIDER 2003 * The authors are Research Associate, Director, and Graduate Assistant, respectively at the Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, Indiana, USA. This study has been prepared within the UNU/WIDER project on The Impact of the WTO Regime on Developing Countries, which is directed by Professor Basudeb Guha-Khasnobis. UNU/WIDER gratefully acknowledges the financial contribution to the project by the Ministry for Foreign Affairs of Finland. Discussion Paper No. 2003/32 OECD Domestic Support and Developing Countries Betina Dimaranan, Thomas Hertel and Roman Keeney 1 April 2003 Abstract An AGE model with detailed farm supply and substitution relationships is used to analyze impacts of OECD domestic support reform on developing economy welfare. Stylized simulations indicate reforms best suited for reducing trade distortions with least impact on farm incomes. Comprehensive reforms result in welfare losses for LDCs and large declines in OECD farm incomes. Shifting from market price support to land-based payments designed to maintain farm incomes results in increased welfare for most developing countries. LDCs should focus on improved market access to OECD economies while permitting said economies to continue domestic support payments not linked to output/variable inputs. Keywords: domestic support, OECD, developing countries, agricultural trade, WTO JEL classification: D58, F13, F14, O19, Q17
Transcript
Copyright � UNU/WIDER 2003* The authors are Research Associate, Director, and Graduate Assistant, respectively at the Center forGlobal Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette,Indiana, USA.This study has been prepared within the UNU/WIDER project on The Impact of the WTO Regime onDeveloping Countries, which is directed by Professor Basudeb Guha-Khasnobis.UNU/WIDER gratefully acknowledges the financial contribution to the project by the Ministry forForeign Affairs of Finland.
Discussion Paper No. 2003/32
OECD Domestic Supportand Developing Countries
Betina Dimaranan, Thomas Herteland Roman Keeney1
April 2003
Abstract
An AGE model with detailed farm supply and substitution relationships is used toanalyze impacts of OECD domestic support reform on developing economy welfare.Stylized simulations indicate reforms best suited for reducing trade distortions with leastimpact on farm incomes. Comprehensive reforms result in welfare losses for LDCs andlarge declines in OECD farm incomes. Shifting from market price support to land-basedpayments designed to maintain farm incomes results in increased welfare for mostdeveloping countries. LDCs should focus on improved market access to OECDeconomies while permitting said economies to continue domestic support payments notlinked to output/variable inputs.
The World Institute for Development Economics Research (WIDER) wasestablished by the United Nations University (UNU) as its first research andtraining centre and started work in Helsinki, Finland in 1985. The Instituteundertakes applied research and policy analysis on structural changesaffecting the developing and transitional economies, provides a forum for theadvocacy of policies leading to robust, equitable and environmentallysustainable growth, and promotes capacity strengthening and training in thefield of economic and social policy making. Work is carried out by staffresearchers and visiting scholars in Helsinki and through networks ofcollaborating scholars and institutions around the world.
UNU World Institute for Development Economics Research (UNU/WIDER)Katajanokanlaituri 6 B, 00160 Helsinki, Finland
Camera-ready typescript prepared by Janis Vehmaan-Kreula at UNU/WIDERPrinted at UNU/WIDER, Helsinki
The views expressed in this publication are those of the author(s). Publication does not implyendorsement by the Institute or the United Nations University, nor by the programme/project sponsors, ofany of the views expressed.
Most studies of global agricultural trade liberalization are primarily focused on marketprice support – that is agricultural support provided indirectly through border measures,either import barriers or export subsidies, designed to boost domestic market prices,relative to world prices (for example, Tyers and Anderson 1992; Martin and Winters1986). In the late 1980s, this form of support accounted for about 75 per cent of totalProducer Support in agriculture in the member countries of the Organization forEconomic Cooperation and Development (OECD) (OECD 2002). Prior to the UruguayRound Agreement on Agriculture (URAA), this was also the only area of agriculturalprotection under negotiation in the international arena. A very important innovation inthe URAA was to put domestic subsidies on the table. More specifically, supportpolicies are placed in ‘boxes’ according to their impact on international trade. Thosepolicies that have ‘no, or at most minimal trade-distorting effects or effects onproduction’ are placed in the green box and are not subjected to reduction requirementsunder the URAA. Those policies that are deemed to be trade distorting are placed in theamber box and are subjected to reductions. However, if the payments are accompaniedby programmes aimed at limiting production, they may be placed in yet a third box, theblue box. As a consequence, they are exempt from the negotiated reductions in support.This third box has since come under scrutiny and there have been proposals to subject itto successive reductions as well – or potentially eliminate this box altogether.
As a result of the URAA, the share of producer support provided by marketinterventions has gradually fallen, so that it now accounts for only two-thirds of totalsupport (OECD 2002). This trend may continue as proposed EU reforms involve furtherefforts to ‘decouple’ support from world prices (The Economist, July 2002).1 The goalof this paper is to assess the likely impact of such decoupling on developing countrywelfare. In the process of making this assessment, we also pay special attention to theimpact of reforms on real farm income in the reforming OECD countries, as the farmlobby is a powerful political force and operates as an important constraint on reformefforts. Due to these dual objectives of the paper, there are necessarily two ratherdistinct parts to our analysis. First, we must assess direct impact of domestic support inthe OECD countries on OECD agriculture – specifically farm incomes, production andsubsequently trade. Then we must assess the impact of these changes on the developingcountries. However, before embarking on this analysis, we first turn to an historicaloverview of domestic support and OECD trade with developing countries.
2 Background on domestic support and developing country trade
2.1 Overview of domestic support in the OECD
The OECD uses the concept of Producer Subsidy Estimates (PSE) as the principalindicator in monitoring and evaluating agricultural policy developments. The PSE is ‘anindicator of the monetary value of gross transfers from consumers and taxpayers toagricultural producers, measured at the farm-gate level, arising from policy measures
1 More recently it appears that France and Germany will oppose such reforms (The Economist, 2003).
2
that support agriculture, regardless of their natures, objectives or impacts on farmproduction or income’. It comprises market price support, payments based on output,payments based on area planted/animal numbers, payments based on historicalentitlements, payments based on input used, payments based on input constraints,payments based on overall farming income, and miscellaneous payments (OECD 2001).The different measures vary in terms of their effects on farm income in the OECDcountries, as well as their effects on trade and hence their impact on the welfare ofdeveloping countries.
Table 1 presents the changes in the overall PSE and its component parts for selectedOECD countries in 1987 and 2000. The PSEs are smallest for Australia and NewZealand. These are largely made up of market price support and variable inputsubsidies. In Australia, for both years, the majority of market price support has beenapplied to grains and milk, while most of the applications of subsidies on variable inputsare applied to meat and meat products. By 2000 most of the PSE had been eliminated inNew Zealand, with large reductions in variable input subsidies in meat and meatproducts. In the case of Japan and Korea, the PSEs have remained relatively unchangedboth in level and in composition. The PSE rates have been highest historically forSwitzerland, but here a fair amount of decoupling has occurred, with the share of marketprice support in the total falling from 82 per cent in 1987 to 59 per cent in 2000.
Table 1Producer subsidy equivalent and components, 1987 and 2000
Per cent share in PSE by support type
Per cent share in PSE by support type
OECD region Year PSE %Marketprice
Output Variableinput
Landbased
Historicalentitlement
Australia 19872000
7.875.56
42.2324.48
02.76
36.5749.66
02.06
05.04
Canada 19872000
35.8419.50
49.8051.22
18.847.12
14.096.43
15.367.63
011.29
EU15 19872000
45.0238.34
85.9258.75
5.515.22
5.496.64
2.7425.42
00.64
Japan 19872000
67.2864.06
90.6891.05
2.562.80
3.954.34
00
00
Korea 19872000
69.4772.56
98.7695.86
00
0.782.45
00
00
New Zealand 19872000
8.870.74
26.7954.43
00
70.3640.31
00
00
Switzerland 19872000
72.9671.38
81.5359.09
1.313.96
8.565.64
6.0611.27
015.86
United States 19872000
27.0121.94
50.8232.01
5.6918.85
14.2113.61
26.607.18
021.51
Source: OECD PSE/CSE Database 2001.
It is in the USA, Canada and the European Union (EU) where sizable cuts in the PSEshow up over this period – although the recent Farm Bill in the USA has reversed thistrend for that country. In the EU, there has been a decided shift in composition ofsupport with the share provided by market price support falling in favour of increased
3
land- and headage-based payments. In Canada, market price support as a PSE share isnearly the same but market price support for grains has been greatly reduced while therehas been a large increase in milk MPS. Most support in Canadian grains is nowprovided via input (including land) and output subsidies, as well as historicalentitlements. In the US, the PSE has been more moderately reduced with a large portionof the reduction coming from the elimination of market price support. In 2000 historicalentitlements had become a much more important component of the PSE in US grains.
This change in the mix of producer support in some of the OECD countries ispotentially quite important. It is also expected to continue – and perhaps accelerate –under a new WTO round. What impact have these historical changes had on worldmarkets? What about prospective changes? For insights on the potential impact ofchanges in the level and mix of domestic support, we turn below to the existingliterature on this topic. But first, let us consider the potential impact that these reformswill have on developing countries. To understand this, we must first examine the tradelinks that will transmit price and quantity changes from OECD countries to developingcountries.
2.2 Overview of developing country trade patterns
Developing countries are an enormously diverse group. Some are net exporters, andsome are net importers of the temperate products that OECD countries tend to protect.Some are closely tied into the OECD markets – by virtue of geography or perhapshistorical trade preferences. Others are more reliant on other developing countries fortheir food supplies and export markets. The strength of the trade links of a developingcountry with the OECD countries will play an important role in the impact of OECDdomestic support reform on the developing country. This section provides an overviewof the trade patterns of developing countries vis-à-vis the OECD countries in agricultureand food products. Data are summarized for the regional and commodity aggregationused in the study provided in Table 2.
Table 3 reports the average trade specialization indices for three decades over the periodranging from 1966 to 1998 for the aggregated regions in this study. Trade specializationindices are calculated as: (X - M) / (X + M) where X are exports and M are imports. Thevalue of the index ranges from -1 for a country which imports, and does not export, aparticular commodity and +1 for a country which is specialized as an exporter of thecommodity. Table 3 separately identifies the aggregated commodity groups –programme commodities,2 livestock and meat products, and other agriculture and foodproducts. Among the developing countries, Argentina has maintained its exportspecialization in programme crops over the period. Economic reforms in Vietnam andIndia have permitted these countries to shift from being moderate net importers to beingnet exporters of programme crops. The net export position of the ASEAN4 region hasseen a decline over the period and Indonesia’s net import position is worsened. China’snet export position has improved. The Middle East/North Africa (MENA) region has
2 The programme commodities referred to in this paper are composed of paddy rice, wheat, cerealgrains, oilseeds, raw sugar, processed rice, and refined sugar. The first four are the crops for which theGTAP database has OECD domestic support data. Processed rice and refined sugar are included sincethese are the traded form of rice and sugar, respectively.
4
Table 2Regional and sectoral aggregation
OECD countriesANZ Australia and New ZealandJapan JapanKorea South KoreaUSA United StatesCanada CanadaMexico MexicoEU15 European UnionEFTA European Free Trade AreaCEU Hungary and PolandTurkey TurkeyDeveloping countriesChina ChinaIndonesia IndonesiaVietnam VietnamASEAN4 Malaysia, Philippines, Singapore, ThailandIndia IndiaRsoAsia Rest of South AsiaArgentina ArgentinaBrazil BrazilRlatAm Rest of Latin AmericaFSU Former Soviet UnionMENA Middle East and North AfricaTanzania TanzaniaZambia ZambiaR_SSA Rest of sub-Saharan AfricaROW Rest of World
Programme commoditiespdrice paddy ricewheat wheatcrsgrns cereal grains nec.oilsds oilseedsrawsgr sugar cane, sugar beetpcrice processed ricerefsgr sugarLivestock and meat productsruminants cattle/sheep, woolnonrumnts animal products nec.rawmilk raw milkrummeat meat: cattle/sheepnrummeat meat products nec.dairy dairy productsOther agriculture and foodothcrops vegetables and fruits, plant-based fibers, other cropsvegoilfat vegetable oils and fatsothprfood other processed foodmnfc manufacturessrvc services
5
Table 3Trade specialization indices: (X-M)/(X+M)
Programme commodities Livestock and meat products Other agriculture and foodregionsRegions
Source: Authors’ calculations from bilateral time series data in GTAP 5 data package.
Note: * The time series trade data for Vietnam starts in 1976 while that for the Former Soviet Union andTanzania start in 1992.
been a consistently strong net importer of programme commodities. Among the OECDcountries, Australia/New Zealand (ANZ) has been consistent in its net export position.The USA and Canada’s net export position has strengthened over the period. The EU15and EFTA have substantially reduced their net imports as a share of total trade, whileJapan and Korea remain consistent net importers of programme commodities over theentire period. Overall, we conclude that increased domestic support for programmecrops appears to have contributed to improvements in the net trade position of theOECD countries in programme crops, at the expense of developing countries.
Turning next to livestock products, we see from Table 3 that China, Argentina andBrazil are net exporters. The specialization indices for these countries, however, havedeclined over the years. On the other hand, the net import positions of India andASEAN4 in these products have diminished markedly. In the OECD countries, the ANZregion stands out as a strong net exporter of livestock and meat products. Japan is astrong net importer and Korea’s net import position has increased over the period. Onthe other hand, the USA, Canada and the EU have seen increases in their tradespecialization indices over the period. Increased domestic support for livestock productsin these countries appear to have contributed to their net export position.
6
Most of the developing countries are consistent net exporters of the aggregate group ofother agriculture and processed food products. Among the OECD countries, Mexico,Turkey and ANZ are net exporters while the other OECD countries in our aggregationare net importers. Thus we have a rough division between temperate products(programme crops and livestock), where OECD domestic support plays an importantrole and where developing countries are largely net importers, and tropical products forwhich developing countries are largely net exporters.3
Focusing next on developing country bilateral trade with the OECD, Table 4 reportsseparately the share of each developing country’s total trade that is specifically withOECD countries. Tanzania, Zambia, and Indonesia rely on the OECD market as adestination for more than three-quarters of their exports of programme commodities. Onthe other end of the scale are Vietnam, Argentina, and the rest of South Asia, of whicheach rely on the OECD market as destination for less than a quarter of their programmecommodity exports. This indicates that a strong net exporter like Argentina competeswith the OECD in third markets for programme commodities. On the import side, theOECD is the source of more than two-thirds of total programme commodity imports ofcountries like China, Indiana and the rest of South Asia and MENA region. For thesecountries, reductions in domestic support for OECD agriculture will mean higher pricedimports. Reforms in OECD market price support may significantly affect the tradepatterns in these countries.
Table 4Share of developing country trade with OECD, 1997
Source: Authors’ calculations from GTAP 5 Data Base.
Notes: * Exports to OECD countries as share of each developing country’s total exports of the commoditygroup. ** Imports from OECD countries as share to each developing country’s total imports of thecommodity group.
3 In order to keep the tables manageable, the other agriculture category also includes food products. Ifthis latter were removed, we would see even more significant net exports from the developingcountries.
7
Even greater dependence on the OECD countries as an import source is exhibited bycountries like China, Indonesia, Vietnam, South Asia, FSU, MENA, Zambia and therest of sub-Saharan Africa in the case of livestock and meat products, with eachimporting more than 80 per cent of their total imports of these commodities from theOECD. Bilateral exports and imports of developing countries for other agriculture andfood commodities are generally less concentrated on the OECD.
Within the group of programme commodities and OECD countries, there is also a greatdeal of heterogeneity regarding the bilateral trading patterns of developing countries.Table 5 reports the shares of bilateral trade of the developing countries with three majorOECD members – Japan, USA and EU – for wheat, a commodity which receivessignificant OECD border protection and domestic support. The first two columns ofdata report each developing country’s share of world trade in 1997 in wheat while thenext six columns report the share of each OECD country as export destination or importsource of each developing country’s total trade in wheat. Developing countries as agroup export 14 per cent and import 54 per cent of total wheat traded in the world. TheUSA and EU each account for a quarter of total world wheat exports. Argentina has an8.8 per cent share of total wheat trade but its export share to the OECD countries is verysmall, Argentina relies on markets other than these three OECD regions. The MENAregion imports 21 per cent of the total wheat traded. Thirty per cent of its wheat importsare sourced from the US and 18 per cent from the EU. The ASEAN4, rest of SouthAsia, and rest of Latin America each account for roughly 4 per cent share of worldwheat imports. The US provides around half of total wheat imported by these countries.
The data examined for wheat in this section of the paper is representative of the broaderpicture of OECD – developing country agricultural trade linkages that are quiteimportant for many products. In the more general case of OECD supported programmecrops and livestock products, many developing countries rely heavily on the OECD fora large share of their imports. These countries may well be hurt by the current trendtowards decoupling domestic support from production decisions as OECD supply pricesare likely to rise as a consequence. On the other hand, those developing countries thatrely heavily on the OECD as an export destination, or that compete with OECDproducts in third markets stand to gain from measures that decouple domestic supportfrom production decisions.
We turn now to a review of the literature analyzing the impact of domestic support onproduction decisions in OECD agriculture.
3 Literature review
The earliest work assessing the impact of different methods of agricultural support onprices, and factor returns in agriculture is that of Floyd (1965). He compared the impactof price supports with output restrictions and mandatory land retirement. He does notconsider the possibility of producer payments based on land use. However, we haveseen above, input-based payments have become increasingly common in recent years.Hertel (1989) develops a series of propositions relating to the impacts of a wider rangeof support measures on production, net exports, employment, land rents and farmincome. He places these on both an equal cost and equal PSE basis for a single product,agricultural sector in the absence of pre-existing support. A few key points emerge from
8
Table 5Shares of developing country trade, 1997
Shares of total developing country tradeShares of totalworld trade Japan USA EUDeveloping
countriesX M X M X M X M
Wheat
China 0.1 3.2 15 0 16 11 39 2
Indonesia 0.0 3.1 14 0 14 1 35 0
Vietnam 0.0 0.0 17 0 25 41 29 0
ASEAN4 0.1 3.7 0 0 0 52 00
India 0.2 1.1 8 0 9 0 22 0
RsoAsia 0.0 4.4 14 0 15 46 36 6
Argentina 8.8 0.0 0 0 0 0 1 0
Brazil 0.1 4.1 0 0 0 0 0 0
RlatAm 0.5 4.9 1 0 1 45 4 10
FSU 2.5 2.3 0 0 0 11 1 12
MENA 1.4 21.1 4 0 5 30 12 18
Tanzania 0.0 0.1 12 0 12 0 30 0
Zambia 0.0 0.0 0 0 0 0 0 0
R_SSA 0.1 3.2 1 0 1 35 2 33
Sugar
China 1.3 2.1 4 1 2 0 5 3
Indonesia 0.3 2.8 33 0 4 1 10 5
Vietnam 0.1 0.0 5 0 20 0 9 1
ASEAN4 9.3 2.3 18 0 9 0 1 6
India 2.0 0.6 2 0 5 0 31 13
RsoAsia 0.6 2.7 2 0 2 0 73 8
Argentina 0.6 0.1 0 0 43 0 0 2
Brazil 14.7 0.0 0 0 7 6 2 12
RlatAm 18.1 3.9 3 0 26 6 18 3
FSU 2.4 9.1 0 0 0 0 3 11
MENA 0.9 15.5 8 0 9 0 29 47
Tanzania 0.1 0.4 6 0 6 0 78 1
Zambia 0.2 0.0 0 1 0 2 97 17
R_SSA 7.6 4.9 4 0 10 0 61 35
Rice
China 6.3 3.2 10 1 6 0 14 0
Indonesia 1.2 1.3 16 0 16 0 39 0
Vietnam 4.8 0.0 0 3 7 0 1 0
ASEAN4 18.6 8.8 5 0 13 1 12 0
India 17.0 0.0 1 6 4 0 13 13
RsoAsia 6.3 2.3 1 0 3 0 4 0
Argentina 3.1 0.1 0 4 1 2 0 1
Brazil 0.1 3.7 5 0 6 1 15 0
RlatAm 7.3 6.8 1 0 1 48 24 1
FSU 0.3 2.0 5 1 5 4 13 8
MENA 1.3 12.9 2 0 2 10 3 4
Tanzania 0.0 0.3 16 0 16 0 40 0
Zambia 0.0 0.0 12 1 17 0 20 30
R_SSA 0.6 4.7 10 2 13 22 24 3Source: Authors’ calculations from GTAP 5 Data Base.
9
this paper. First of all, subsidies on variable inputs that substitute for fixed factors (forexample, land) in agriculture have a greater impact on output, and hence trade, than doequal cost output subsidies. Such variable input subsidies also moderate the share ofproducer support that accrues to land and other fixed factors. On the other hand,subsidies to land, such as the per hectare payments currently made in the EU, have amore modest effect on output, while leading to higher land rents than under an equal costoutput subsidy. Finally, when compared to an output subsidy of equal cost, exportsubsidies have a larger impact on exports, agricultural production, employment, and landrents, provided the elasticity of export demand exceeds the domestic demand elasticity.
Subsequent work in this area has been largely computational in nature. Abler and Shortle(1992) focus their attention on the relationship between chemical restrictions andexisting farm programmes in the US and the EU. They find that unilateral restrictions onchemical usage benefit US farmers, while leading to losses on the part of EU producers.Gunter et al. (1996) focus on input market interventions as well, evaluating their impacton competing policy goals in the context of a three region, US-EU-ROW, partialequilibrium model of wheat markets. Of special interest for the present paper is therecently completed OECD (2001) report on ‘Market Effects of Crop Support Measures’.In this report, the authors compare the impacts of a wide range of producer supportacross OECD countries. They find that the movement from market price support andoutput subsidies to land-based payments is a ‘win-win’ scenario in most countries – withfarm income rising and world price impacts of support falling.4 From the point of viewof this paper, this suggests an interesting possibility, namely that re-instrumentation ofproducer support for agriculture in OECD countries could conceivably maintain OECDfarm incomes, while contributing to enhanced welfare on the part of developing countryexporters. This hypothesis will be explored in greater detail below.
In a separate study, also undertaken by the OECD Agriculture Directorate, Thompsonand Smith (2002) study the impact of further agricultural trade reforms on developingcountries. They use two modelling frameworks. The OECD AgLink model is used toexamine the impacts of reductions in market price support, while the GTAP model isused to examine the impacts of cuts in both market price support and direct payments toproducers. They look at relatively broad groups of developing countries, and they do notconsider more elaborate reforms in which the mix of measures is changed in an attemptto maintain farm incomes.
In contrast, Frandsen et al. (2002) use a modified version of the GTAP model toexamine the impact of further decoupling of domestic support in the EU. Theiremphasis is on the budgetary and macro-economic effects of these policy reformsamong OECD countries. They argue that further decoupling of EU agricultural policieswould reduce budgetary exposure in the EU as well as bringing it into compliance withpotentially stricter WTO disciplines on domestic support. They also find rathersubstantial changes in world prices – particularly for meat products, although they donot examine the issue of overall developing country welfare explicitly, and they restrictthemselves to EU reforms.
4 One cautionary note – as anticipated in the results of Hertel (1989) is that a shift towards variableinput subsidies could have the opposite effect – with larger world price impacts and smaller farm-income benefits
10
The goal of this paper is to assess the impact of changes in both the mix and the level ofdomestic support in OECD countries on the welfare of farm households in the OECDand on the national welfare of developing countries. Therefore, it is not enough to saythat world prices will rise or they will fall. The welfare impacts on developing countrieswill depend on whether they are net exporters or net importers of protected products. Itwill also depend on the bilateral trade patterns discussed in Section 2. Are they closelytied into the OECD markets in which these changes occur? In short, we need a globaltrade model with bilateral trade flows explicitly treated. One such framework is offeredby the Global Trade Analysis Project (GTAP) database and associated models, used bya number of the preceding studies.
Since the early 1990s, there has been a large number of global, general equilibrium,analyses of trade liberalization – some of which include domestic support (these includeFrancois et al. 1996; Hertel et al. 1996; Harrison et al. 1996; Anderson et al. 1999;Elbehri et al. 1999; Hertel and Martin 1999; Anderson et al. 2001; Rae and Strutt 2002).Most of these studies are based on the GTAP database and modelling framework.However, the GTAP database has not been particularly well-suited to the analysis ofdomestic support issues. Versions 1–4 of the GTAP database treated all domesticsupport as an output subsidy. Version 5 introduced a first-cut disaggregation of supportacross inputs (Dimaranan 2002), but it still suffers from some important limitations(Gehlhar and Nelson 2001; Frandsen et al. 2001). Furthermore, the standard GTAPmodel is not well-suited to analysis of domestic support issues, due to its relativelysimplistic treatment of factor markets. One contribution of the present paper is toaddress these limitations. Thus we now turn to the issue of model design.
4 Methodology
4.1 Model design
For purposes of this study, we have constructed a special purpose version of the GTAPdatabase and model, designed to make it more appropriate for the analysis of domesticsupport. We adopt, as our starting point, the general framework proposed by the OECD(2001) in which factor demand and supply relations play a central role. The mostvaluable contribution of this report resides in the annexes, where extensive literaturereviews are available for the EU and for North America. The authors provide centralparameter values for the key elasticities of substitution, as well as for factor supplyelasticities (see tables A1.3 and A1.4 of OECD 2001). We have restructured the GTAPmodel in order to take advantage of this information and it is to these features that wenow turn.
We begin by segmenting the factor markets for labour and capital between agricultureand non-agriculture. A key parameter in the OECD analysis is the elasticity of factorsupply for farm-owned inputs. The values of these parameters, as well as the ranges,proposed by the OECD are reported in Table 6. Note that these values are less than one,which is a sharp contrast to the usual assumption of perfect factor mobility used in mostCGE analyses. This means that commodity supply is also less responsive, and more ofthe benefits of farm subsidies (or losses from their elimination) will accrue to farmhouseholds.
11
Table 6Factor supply and substitution elasticities adapted from OECD (2001)*
Notes: * Data ranges in parentheses. ** The data provided in OECD (2001) cover only Japan,USA, Canada, Mexico, EU, and Switzerland. We adapted data Canada’s data for Australia/NewZealand, Japan’s data for Korea, and Switzerland’s data for EFTA. Data for Mexico wasassigned to the CEU (Hungary and Poland), Turkey and all the developing countries.
The OECD report also attempts to come up with supply elasticities for purchased inputs.However, there is little econometric evidence to draw on here. One advantage of thegeneral equilibrium framework is that these commodity supply responses areendogenously determined – as a function of the factor market assumptions as well as thecost structure of the industry. Therefore, we dispense with the OECD estimates of inputsupply for fertilizer and other purchased inputs. The supply prices for the 18 differentintermediate inputs are endogenous in the model and determined by the interaction ofsupply and demand in each of these markets.
12
On the factor demand side, we employ a nested-CES production function which can becalibrated to the three key elasticities of substitution available from the OECD report(Table 6). Specifically, we postulate that output is a CES composite of two inputaggregates. The first of these is a purchased input aggregate, while the second is a value-added aggregate. The individual inputs in each of these groups are assumed to beseparable from one another – with a common elasticity of substitution. The purchasedinput and value-added aggregates are themselves each a CES function of individual farminputs. This gives us a total of three CES substitution parameters. They are calibrated tothe OECD central values for the Allen partial elasticities of substitution between: (i) landand other farm-owned inputs, (ii) land and purchased inputs, and (iii) among purchasedinputs. These values are reported in Table 6 for the OECD countries covered in thereport. These parameters are not critical for our analysis of the non-OECD impacts, sincedomestic policies in these countries are unchanged in our simulations. Accordingly, wesimply set these parameter values equal to those from Mexico for all non-OECDcountries in the model.
Given our interest in tracking real farm income and the overall measure of support forOECD agriculture, we also add some additional equations to the model to determinethese variables. Real farm income is based on payments to endowments in the farmsector, adjusted for depreciation and the farm sector’s share of national net taxes. Toobtain real farm income, we deflate this by the regional household’s price index which iscomputed in the standard GTAP model. In some simulations, real farm income is treatedas exogenous, and a policy instrument is endogenized in order to maintain this targetlevel of income.
The computation of PSEs in the GTAP model is complicated by the fact that tradedcommodities are differentiated by origin. So the model tracks bilateral trade and there isno unique world price. Therefore, the domestic-world price gap is measured as a trade-weighted combination of bilateral import and export prices. In the case of market pricesupport, this price gap is applied to output in order to compute the change in PSEassociated with a given policy change. In some simulations, the PSE – either at thecommodity or sector level – is exogenized and a policy instrument is endogenized tomaintain this pre-specified level of support.
Finally, given the importance of the trade elasticities to our analysis, we haveincorporated recent estimates, implemented at the disaggregated GTAP level, based onthe methodology outlined in Hummels (1999). Here, he uses detailed trade, tariff andtransport cost data for a variety of importing countries in North and South America toestimate a differentiated products model of import demand. The variation in bilateraltransport costs permits him to get quite precise estimates of these parameters – in sharpcontrast to much of the earlier work in this area.
The remainder of the model follows the standard GTAP framework, with sectorsproducing output under perfect competition and constant returns to scale. Consumerdemands are modeled using the non-homothetic, CDE functional form, calibrated toestimates of price and income elasticities of demand. Bilateral trade flows are modeledusing the common, Armington approach under which products are differentiated byorigin. Bilateral transport costs between countries are explicitly modeled, and a global‘bank’ serves to close the model with respect to global savings and investment.
13
4.2 Data and aggregation
The study uses an aggregation of a revised version of the GTAP 5 database (Dimarananand McDougall 2002). In the GTAP 5 database, all the different components of OECDPSE data except for market price support are distributed into four classifications ofdomestic support namely: output subsidies intermediate input subsidies, land-basedpayments and capital based payments (Jensen 2002). In contrast to GTAP 5, the land-based payments were revised to separately handle payments on historical entitlements.Their effect is now neutral across programme commodities. The region and sectoraggregation of the GTAP database used in the study is a laid out in Table 2.
4.3 Experimental design
Five sets of simulations are used in this paper to analyze the impacts of changes inOECD domestic support on developing regions. The experimental design is outlined inthe list below.
Experimental design
(a) Stylized shocks Perturbations equivalent to a one per cent increase in the PSE,assuming no initial subsidies applied to each of market pricesupport, output subsidy, input subsidy, and area payments forwheat in the EU (Table 7).
(b) Interactions withexisting subsidies
Land subsidy, variable input subsidy, output subsidy, or marketprice support is allowed to adjust to maintain when a one percent shock is applied to the EU15 PSE (Table 7).
(c) Policy reform and re-instrumentation for EUwheat
EU wheat land subsidy is allowed to adjust to maintain aconstant real farm income condition when market price supportis reduced by 50 per cent (Tables 8-9).
(d) 50 per cent cuts inOECD domesticSupport
Comprehensive reform of domestic support in OECD for allcountries and all commodities: 50 per cent cuts in all domesticsupport instruments (Tables 10-12).
(e) 50 per cent cuts inOECD market pricesupport with re-instrumentation
Comprehensive reform of market price support, including 50 percent cuts in tariffs and export subsidies, with a compensatingincrease in payments to land, designed to stabilize real farmincome in each OECD country (Tables 13-15).
The first set of simulations involves shocking each type of domestic support and marketprice support by the PSE equivalent of a one per cent increase in market price support.These equal PSE conditions are derived in Hertel (1989) under the assumption of zeroinitial distortion. These results are the key to understanding the domestic support model,as the equal PSE condition as derived here highlights the relative responsiveness of keyindicators to equivalent changes in support measures.
The second set of simulations builds on the first by enforcing an actual equal PSEcondition on the model solution. This is done by solving the model in response to a oneper cent shock to the EU15 PSE, with the change in a particular support instrument beingconsidered made endogenous. The results for these simulations highlight the importanceof interactions of changes in support instruments as well as the importance of pre-existing tax/subsidy levels in a reform process that changes the composition of support.
14
Table 7Equal PSE comparison across alternative support instruments
Equal PSE stylized:+1% shock
Equal PSE actual:+1% PSE shock
EU15 indicator Land Output MPS Input Land Output MPS Input
Real farm income 0.98 0.10 0.04 -0.01 0.09 0.06 0.04 -0.01
Source: Authors’ simulations.
Notes: a Varies by importing or exporting region; * Zero initial distortion is assumed.
Table 8Implications of 50 per cent reduction in market price support for EU15 wheat, with
re-instrumentation
EU15 variable Per cent change
Change in area payments -8.6
Land rents 0.3
Wheat acreage planted 0.0
Labour use -3.4
Capital use -3.3
Output price -0.7
Output quantity -3.3
Export price 0.6
Export quantity -7.5
World price 0.4
Equivalent variation US$ million
EU15 187.8
OECD-FSU aggregate 246.7
Developing region aggregate -69.0
Source: Authors’ simulations.
15
Table 9Developing region welfare: EU15 wheat market price support reform in US$ millions
(percentage change in parentheses)
Equivalent variation Terms of trade components
Region* TotalAlloc.
efficiency I-S effect TOTWorldprice
Exportprice
Importprice
LDC total(-0.67)
-65.5 -9.6 -0.8 -55.2 -34.7 12.7 -33.2
China(-0.91)
-4.8(-0.001)
-2.8 -0.2-1.8
(-0.001)-2.1
(-0.001)2.5
(0.001)-2.2
(0.001)
Indonesia(-1.00)
-3.1(-0.002)
-0.1 0.0 -3.0(-0.005)
-3.4(-0.006)
0.7(0.001)
-0.3(0.000)
Vietnam(-1.00)
-0.0(-0.000)
-0.0 0.0 -0.0(-0.000)
-0.1(-0.002)
0.1(0.003)
-0.0(0.000)
ASEAN4(-0.97)
-3.9(-0.001)
-0.9 -0.1 -3.0(-0.001)
-3.6(-0.001)
2.5(0.001)
-1.9(0.001)
India(-0.83)
-0.4(-0.000)
-0.2 0.0 -0.2(-0.000)
-1.8(-0.003)
1.3(0.002)
0.3(-0.000)
RsoAsia(-1.00)
-6.3(-0.005)
-1.1 -0.2 -5.0(-0.019)
-4.2(-0.016)
0.1(0.000)
-1.0(0.004)
Argentina(1.00)
7.0(0.002)
0.8 0.3 5.9(0.020)
5.8(0.020)
0.3(0.001)
-0.2(0.001)
Brazil(-0.96)
-3.6(-0.001)
-1.0 -0.1 -2.5(-0.003)
-3.7(-0.005)
0.8(0.001)
0.4(-0.001)
RlatAmer(-0.86)
-10.1(-0.002)
-1.6 -0.1 -8.3(-0.006)
-6.5(-0.004)
4.4(0.003)
-6.2(0.004)
MENA(-0.88)
-29.6(-0.005)
-2.0 0.0 -27.6(-0.012)
-11.7(-0.005)
0.8(0.000)
-16.7(0.007)
Tanzania(-1.00)
-0.1(-0.002)
-0.0 0.0 -0.1(-0.005)
-0.1(-0.007)
0.0(0.001)
0.0(-0.000)
Zambia(0.76)
0.0(0.000)
0.0 0.0 0.0(0.001)
-0.0(-0.000)
-0.0(-0.000)
0.0(-0.002)
R_SSA(-0.94)
-10.4(-0.004)
-0.8 -0.1 -9.6(-0.010)
-3.2(-0.003)
-0.8(-0.001)
-5.5(-0.006)
Source: Author’s simulations.
Note: * Specialization indices in italics.
16
Table 10Change in average world prices due to comprehensive OECD domestic support reform
(50 per cent reduction) (percentage change in parentheses)
Contribution by tax/subsidy to world price change
CommodityWorld price
change OutputIntermediate
input Land Capital
pdrice 0.26 0.12 0.34 0.05 -0.23
wheat 4.91 1.03 1.68 1.11 1.09
crsgrns 5.5 1.42 1.79 1.02 1.27
oilsds 3.53 0.92 1.21 0.79 0.6
rawsgr -0.58 0.09 0.14 -0.33 -0.48
othcrops -1.5 -0.01 -0.03 -0.69 -0.77
ruminants 4.3 0.48 0.95 -0.38 3.25
nonrumnts 0.54 0.26 0.45 -0.14 -0.02
rawmilk 0.21 0.14 0.81 -0.33 -0.4
pcrice 0.27 0.13 0.12 0.06 -0.03
vegoilfat 0.97 0.2 0.34 0.24 0.2
refsgr -0.06 0.05 0.06 -0.03 -0.15
rummeat 2.21 0.31 0.56 -0.11 1.44
nrummeat 0.43 0.17 0.28 -0.06 0.04
dairy -0.19 0.14 0.36 -0.27 -0.43
othprfood 0.22 0.06 0.11 0.07 -0.03
mnfc 0.12 0.01 0 0.1 0.01
srvc 0.11 0.01 0 0.1 -0.01
Source: Authors’ simulation.
17
Table 11Developing region welfare changes: domestic support reform in US$ millions
(percentage in parentheses)
Equivalent variation Terms of trade components
Region TotalAlloc.
efficiency I-S effect TOTWorldprice
Exportprice
Importprice
China-69.1
(-0.009)-69.6 -18.0
18.5(0.005)
-51.8(-0.015)
137.1(0.039)
-66.8(0.019)
Indonesia-13.6
(-0.007)0.8 -1.9
-12.4(-0.021)
-54.5(-0.095)
35.5(0.062)
6.6
(-0.012)
Vietnam-8.2
(-0.042)-1.9 0.3
-6.6(-0.071)
-10.0(-0.107)
5.8(0.062)
-2.4(0.026)
ASEAN4-15.2
(-0.004)4.9 -4.3
-15.9(-0.004)
-47.4(-0.013)
113.4(0.031)
-81.9(0.022)
India35.9
(0.010)15.2 -2.1
22.8(0.049)
-22.9(-0.049)
38.6(0.083)
7.1(-0.015)
RsoAsia-44.2
(-0.037)-3.3 -1.2
-39.7(-0.149)
-57.2(-0.214)
17.2(0.064)
0.3(-0.001)
Argentina157.3
(0.053)26.2 10.6
120.5(0.428)
183.1(0.653)
-53.1(-0.189)
-9.5(0.034)
Brazil200.2
(0.029)73.3 31.9
94.9(0.173)
1.1(0.002)
88.5(0.161)
5.3(-0.010)
RlatAmer-214.3
(-0.050)-29.9 -1.0
-183.4(-0.135)
-244.7(-0.180)
101.8(0.075)
-40.5(0.030)
MENA-270.1
(-0.045)-50.6 -1.8
-217.7(-0.091)
-315.9(-0.132)
83.1(0.035)
15.1(-0.006)
Tanzania-7.0
(-0.111)-1.2 -1.0
-4.9(-0.420)
-7.1(-0.608)
1.8(0.154)
0.4(-0.035)
Zambia0.0
(0.000)0.2 0.0
-0.3(-0.017)
-1.4(-0.103)
0.4(0.031)
0.7(-0.055)
R_SSA-126.1
(-0.424)-16.0 -2.1
-108.0(-0.120)
-149.7(-0.166)
31.1(0.034)
10.6(-0.012)
ROW17.1
(0.002)27.7 -1.1
-9.4(-0.001)
-221.4(-0.029)
285.9(0.037)
-73.9(0.010)
LDC Total -357.3 -24.2 8.4 -341.6 -999.7 887.0 -228.9
Source: Authors’ simulations.
18
Table 12Terms of trade welfare contribution decomposed by region and commodity: comprehensive 50 per cent reduction in OECD domestic support in US$ millions
World price effects by region
Com. China Indon. Vnam ASEAN Ind. RSoA Arg. BrazRLatAmer MENA Tanz. Zamb. R_SSA ROW
The third set of experiments detailed in the list involves a 50 per cent liberalization ofborder measures relating to wheat in the EU15 while allowing area payments to adjustto maintain the real farm income level in the EU15. This simulation provides the firstinsight into changes in model variables that result from a politically feasible reformscenario. In addition, the focus on reform in a single region-commodity pair provides agood starting point for examining the mechanisms underlying the welfare impacts ondeveloping regions occurring from OECD reforms.
The final two sets of experiments consider more comprehensive reforms in which firstdomestic support and then market price support are cut by 50 per cent in all OECDcountries. In the second simulation, domestic support is endogenized to offset theadverse impacts of cuts in market price support for OECD farm incomes.
4.4 Differential impacts of alternative farm support policies
As described above, these experiments are useful because they provide comparativestatic-based insights into what we can expect the changes in key model variables to be.The shocks applied here are from those derived in Hertel (1989) as equal PSE shocks,based on the assumption that there are no distortions in place initially. This exercise alsoserves as a way to validate the model in light of the algebraically derived expectedresults – setting the stage for more complex simulations. Results for the stylized PSEshocks are given in the first four columns of Table 7.
The results shown above conform to those predicted in Hertel (1989) as well as to theempirical results presented in the OECD (2001). An equal PSE increase to the subsidyon variable inputs has the largest effect on wheat output, exports, and prices, as farmersare encouraged to boost yields in the wake of cheaper land-substituting inputs. Withland becoming less scarce, returns to land decline under this scenario, thereforecontributing negatively to real farm income. This type of ‘subsidy’ does not benefitfarmers at all!
In contrast, subsidy payments to land used in production of wheat have the smallesteffects on output, exports, and price of wheat. With an inelastic supply of land to wheatproduction, a substantial portion of the subsidy is capitalized in higher land values, andfarm income is increased substantially. This result is reinforced by the addition of a set-aside requirement. Here, we apply a simple rule of proportionality. In the base year, theset aside requirement was 10 per cent, so a 5 per cent increase in the land subsidy wouldbe accompanied by a 0.05 * 0.10 = 0.5 per cent increase in set aside.
The results for an output subsidy as compared to market price support show that for thismodel as is seen in the results for OECD (2001), output subsidies have a larger effect onoutput, producer prices, and farm income (through land rents) than does market pricesupport. In our model, the output subsidy is also more trade-distorting, which flies in theface of simple theoretical results. This is due to the role of ‘own-use’ in the GTAPmodel. Sectors tend to purchase their own output as an input. Under the output subsidy,the cost of these ‘inputs’ falls, whereas it rises under the export subsidy.
23
4.5 Interaction with existing subsidies
The stylized shock results assumed that there were no pre-existing subsidies in place inorder to highlight relative responsiveness to changes in different types of support.However, given the initial distortions in place it is necessary to impose an equal PSEcondition on the model solution to arrive at actual equal PSE model results. Results forsimulations carried out under this condition are presented in the right hand four columnsof Table 7.
The first variable reported in Table 7 gives the initial ad valorem rate of the tax orsubsidy for each instrument. From this, it is clear that, in 1997, there was a verysubstantial initial subsidy on land in EU wheat production. This means that the impactof marginal changes in spending on the land subsidy will be blunted by the fact that agiven per unit subsidy will now represent a much smaller portion of the rental price ofland. This point is made forcefully by the OECD (2001) in their analysis of crop supportpolicies. This set of simulation results highlights this point, by evoking an actual equalPSE response in each type of instrument that is very different than that observed underthe zero initial distortion assumption.
The results for a land subsidy under the actual equal PSE simulation shows the changein land subsidy necessary to increase the PSE by one per cent and the associatedimpacts of increasing support via area payments. Note that land returns in wheat andfarm income rise, but not by nearly as much as would have been expected based on theresults in column one of Table 7. Increasing the PSE in the EU15 by one per cent via anoutput subsidy boosts land rents by nearly two-thirds as much as the land subsidy case –whereas the same factor of proportion in column two of Table 7 was less than one tenth.This is due to the fact that the initial level of output subsidy for EU wheat is negligible.Clearly the initial level of support matters.
4.6 Policy re-instrumentation
A primary obstacle to reducing agricultural support in OECD agriculture is the adverseimpact on farm incomes. Given the differential impact of the various methods ofsupport used in OECD countries, as illustrated in Table 7, there appears to be scope forre-instrumentation of support. This point is made quite clearly by Dewbre et al. (2001),who show that market price support is a relatively inefficient means of transferringincome to farmers and furthermore, that it does so at the expense of relatively largedistortions in world markets. They show that, in contrast, land-based payments arehighly effective at transferring income to farmers, while reducing world market priceimpacts of OECD agricultural policies. Therefore, we turn next to a simulation in whichmarket price support for EU wheat production is further reduced (by 50 per cent from1997 levels), yet farm income is maintained at current levels by increased land-basedsubsidies. This simulation is really just an extension of the kind of reform that the EUhas been undertaking over the past decade.
Reducing MPS by 50 per cent and maintaining farm incomes via area payments resultsin an 8.6 percentage point increase in the power of the ad valorem land subsidy (treatedas a negative tax in the model). Note that we have not increased the set-asiderequirement in this case, since output falls due to the reduction in domestic prices. Theincreased subsidy to wheat land results in increased returns to land employed in the
24
wheat sector, which in turn attracts more land to this activity. With overall productiondeclining, this policy leads to a more extensive form of wheat production, with a declinein the use of labour, capital and purchased inputs per hectare of land.
The decline in wheat production and increase in consumption due to lower domesticprices causes wheat exports from the EU15 fall. This is further reinforced by thereduction in export subsidies for wheat. Thus the export price of EU wheat rises. Thisre-instrumentation leads to an increase in efficiency in the EU economy and asubsequent welfare gain of US$188 million. With the exception of Argentina andZambia, the developing countries lose from higher wheat prices. The aggregate welfareloss to developing countries totals US$65 million in this case. Given the goal of thispaper, a more refined examination of developing country welfare impacts in the wake ofthe simulated reform is the appropriate place to turn our discussion.
4.7 Impacts on developing countries
The developing country impacts of the EU wheat reform summarized in Table 8 aredecomposed by region and welfare contribution in Table 9. Here, we follow theapproach of Huff and Hertel (1996) whereby regional welfare can be explained byallocative efficiency effects and the terms of trade effects. The allocative efficiencyeffects are due to second-best effects where a country benefits positively from increasedactivity in industries that are taxed and negatively from the expansion of subsidizedindustries. The terms of trade effects come from changes in a country’s export pricesrelative to changes in its import prices. A country benefits positively from an increase inits export prices and is negatively impacted by a net increase in the prices of goods thatit imports. As noted previously, since the developing country impacts of OECD reformare transmitted through international markets, it is hardly surprising that the resultingchange in the terms of trade for these countries (TOT in Table 9) account for the bulk ofthe developing country losses. Furthermore, with the exception of Argentina andZambia, all of the developing countries are made worse off due to the EU15 wheatreform.
It is challenging to sort out the impact of changes in export and import prices ofdifferent commodities in order to explain why a given country experiences a terms oftrade gain or loss. A helpful approach to decomposing the terms of trade effects isprovided by McDougall (1993) who decomposes the percentage change in the terms oftrade for a given region into three separate effects – the world price effect, the exportprice effect and the import price effect:
� �−=i i
MirMriXir
Xrir pSpStot effect tradeof erms t
) )((� −−=i
WWiMri
Xri ppSS effect price ld wor
)(� −+i
WiXirXr
i ppS effect priceexport
)(� −−i
WiMirMri ppS effect priceimport
25
The world price effect equals the sum over all traded commodities of the product of acountry’s net trade share (the difference between export and import shares forcommodity i), (SiXr – SiMr), and the change in the price of i (for example, wheat), Pwirelative to an index of average world prices for all products, Pw. (Lower case variablesdenote percentage change so the difference in these two price changes represents thepercentage change in the price ratio.) The world price effect is positive in the case of anet exporter of a commodity for which EU reform means higher world prices. However,from Table 9 (see specialization indexes in parentheses below each country) we knowthat most of these developing countries are net importers of wheat. Therefore thiscomponent contributes negatively to their welfare.
The MENA region suffers the worst absolute and relative (percentage) deterioration interms of trade due to the world price effect, owing to MENA’s heavy reliance onimports of wheat. Examining the entries in the world price effect column of Table 9, wesee that Argentina, which is a substantial net exporter of wheat, is the one countrywhich experiences a welfare gain from the higher world wheat prices. (Zambia is also asmall net exporter of wheat in our base period, but the gain on this commodity is offsetby losses on more important export commodities.)
The second component in the terms of trade decomposition is the export price effectwhich is the sum of export share-weighted relative price changes where the relativeprice change is the ratio of the exporter’s price for commodity i, PXir, relative to theworldwide average price for commodity i, Pwi. Of course, if these commodities areperfect substitutes, then this effect disappears since the two prices will not differ in thecase of a homogeneous commodity. The degree to which the two prices can diverge isinfluenced by the degree of product differentiation in the market for commodity i. Thereis product differentiation in all commodities in this model since the Armington tradestructure ensures that wheat produced in one country is differentiated from wheatproduced in another. The extent of differentiation is based on a new set of econometricestimates undertaken at the GTAP level of aggregation, following the work of Hummels(1999).
The export price effects in Table 9 are uniformly positive, with the exception of Zambiaand the rest of sub-Saharan Africa. These positive entries reflect the fact that increasedEU imports of wheat result in higher EU exports of other products, and thereby lowerEU export prices. Since the world average price for all goods is a weighted average ofall export prices, most non-EU export prices rise, relative to the average.
The import price component of the terms of trade decomposition is the mirror image ofthe export price effect and refers to the import share-weighted change in the country-specific import price index, PMir, relative to the average world price index, Pwi.Developing countries tend to receive subsidized imports from the EU and so it is hardlysurprising that elimination of these subsidies results in higher average prices forcomposite wheat imports. This effect is particularly important for MENA, rest of LatinAmerica, and rest of sub-Saharan Africa.
The final column in Table 9 reports a residual component of the developing countrywelfare impacts that we have also included in the TOT total. This has to do with
26
changes in the price of capital goods used for investment purposes. It is relatively minorand will not be discussed further here.5
4.8 Analyzing the impact of comprehensive OECD agricultural reforms ondeveloping countries
Having worked through the basic mechanisms by which domestic support andprotection of OECD markets will affect the developing countries, we now ‘scale up’ thisanalysis of one specific commodity to the global level by examining the combinedimpact of cuts in support for all agricultural products in all OECD countries. We beginby examining the impact of a 50 per cent cut in domestic support, then turn to anexperiment akin to the one discussed before whereby market price support is cut by 50per cent, while domestic support in the form of area payments rises to stabilize OECDfarm incomes.
4.9 Cutting domestic support in the OECD
The first column of Table 10 reports the average world price impacts of cuttingdomestic support for all agricultural commodities in the OECD by 50 per cent. It isimmediately clear that domestic support policies have the strongest impact onprogramme crops and ruminant livestock (primarily beef). These are the commoditieswhere the world price increases are greatest. Sugar and dairy, where the bulk ofprotection remains at the border, actually shows small price declines, as land and labourshifts out of programme crops into other activities. This also causes other crop prices tofall as well.
The remaining columns of Table 10 decompose the total world price effect by type ofdomestic support policy instrument, including output subsidies, intermediate inputsubsidies, land-based payments and capital subsidies (including livestock-basedpayments). Despite the importance of land-based payments for programme crops in theEU and USA, it is the intermediate input subsidies that contribute most to the worldprice effects for these crops stemming from domestic support policies in the OECD. Forexample, 1.7 per cent of the 4.9 per cent increase in the world price of wheat followingthis cut in domestic support is attributed to the cut in intermediate input subsidies. Thisis due to the fact that they are both important in the overall mix of support (see Table 1)as well as highly distorting of world trade, as demonstrated in Table 7. In the case of thestrong increase in the price of ruminant meat, this is largely due to the subsidies onanimal numbers (capital subsidy).6
The impact of this domestic support reduction scenario on developing country welfare isreported in the first column of Table 11. As can be seen from this Table, developingcountries as a group lose from this cut in OECD domestic support. The notable
5 For those familiar with GTAP, this is the component of the welfare decomposition that refers to thepurchases of savings from the ‘global bank’ and the sales of investment goods to that same entity. Seethe technical paper by Huff and Hertel (1996) for further discussion and interpretation of this term.
6 These results can be compared roughly to those of Rae and Strutt (2002) by noting that they omit theland and capital-based payments from their domestic support scenario, arguing that these are largely‘blue box payments’ and therefore exempt from cuts under the Uruguay Round agreement.
27
exceptions are Argentina, Brazil and India. The next two columns of this tabledecompose these welfare effects into their allocative efficiency and terms of tradecomponents. As with the previous wheat example, the bulk of the developing countrylosses are due to the deterioration of their terms of trade. The only case where theallocative efficiency effect dominates is for China. This is largely driven by theinteraction between reduced oilseed imports from the USA, interacting with a very highpre-WTO accession tariff on these imports. That tariff has since been dramaticallyreduced as part of China’s WTO accession process (Ianchovichina and Martin 2002) sothis effect is no longer empirically relevant.
As before, we can decompose the terms of trade effect into its component parts to obtainsome further insight into the source of the developing country losses. This is done in thesubsequent three columns of Table 11. Note that the world price effects are dominant,and negative, followed in magnitude by the export price effects which are positive fordeveloping countries as a group. The import price effects are negative, and considerablysmaller in absolute value.
Table 12 breaks out the world price effects by commodity and region. Recall that theworld price effect is positive when the price rises and the country is a net exporter andnegative when it is a net importer. For a world price decline, it is precisely the opposite.From Table 10, recall that the world price rises were most dramatic for the programmecrops and for ruminant meats, while the biggest price decline is for other crops.Furthermore, recall from Table 3 that developing countries tend to be net importers ofprogramme crops and livestock products, and net exporters of other crops. Therefore, itis not surprising that the largest losses are for wheat, coarse grains, ruminant products(net importers with a world price rise) and for other crops (net exporters with adeclining world price). From the point of view of an individual region/country, MENAand rest of Latin America are among the hardest hit by these effects.
Recall, however, that our analytical framework takes into account the differentiation ofproducts by country of origin. So the export price effect can potentially offset orreinforce the world price effect, depending on whether developing country export pricesrise or fall, relative to the world average. The last set of columns in Table 12 report theexport, import and total TOT price effects, by commodity for developing countries as agroup. Here, it can be seen that the product differentiation aspect of the analysis furtherreinforces the adverse impacts on developing countries for wheat, coarse grains,oilseeds, and ruminant products. However, in the case of other crops, which are quitehighly differentiated, the rise in developing country export prices, relative to the worldaverage, generates an overall gain. Developing countries also benefit overall fromdevelopments in the global markets for manufactures and services.
In addition to the losses incurred by developing countries from the cuts to domesticsupport in the OECD countries, there are substantial declines in OECD farm incomes.The largest decline is in the EU15 (-16 per cent), followed by EFTA (-13 per cent), thenUSA (-5 per cent) and Canada (-3.5 per cent). The losses in most other OECD countriesare under one per cent, due to relatively more reliance on border measures (Japan andKorea – see Table 1) or lower levels of support (Australia and Canada). From a politicaleconomy point of view, this kind of reform looks like a difficult one to sell. Thereforewe turn to an alternative type of comprehensive reform. This builds on the idea of re-instrumentation that was developed in the first part of the paper.
28
4.10 Re-instrumentation of agricultural support in the OECD
In this section of the paper we simulate an alternative type of comprehensive, OECDreform focusing on reductions in market price support. Specifically, tariffs and exportsubsidy rates in the OECD countries are cut by 50 per cent. Domestic support is actuallypermitted to increase in order to compensate producers for the resulting loss in income.As with our EU wheat example above, we use the land-based payments to compensateproducers, since they are the most efficient and least trade-distorting of the instrumentscurrently in use.
Table 13 reports the world price effects of the re-instrumentation experiment. The firstcolumn reports the total effect, while the subsequent columns break this total into theparts attributable to tariffs in the major OECD markets, as well as export subsidies (EUand other OECD). The first thing to note is that the world price effects on programmecrops and ruminant products are far more modest than those following the domesticsupport experiment. In general, the average world price of crops rises, while the averageworld price of livestock products falls. The largest contributor to the higher rice pricesis the Japanese tariff cut. In the case of wheat prices, EU export subsidies, followed byJapanese tariffs, are the largest contributions to the increase. The situation is similar forcoarse grains, where the majority of the world price impact is traced back to theelimination of EU export subsidies. The average world farm gate price of sugar risesdue to cuts in the EU and US import tariffs. Meat and dairy prices world-wide areheavily influenced by the EU tariff cuts. With a large share of the world’s output in theEU, lower prices in that market contribute to a decline in the world average price.Finally, in the case of other food products, the ‘other’ OECD countries tariffs appear toplay the largest role.
Table 14 reports the welfare impacts of the re-instrumentation experiment. Now we seethat, in sharp contrast to the domestic support experiment, most developing countriesgain from the liberalization. Only China, ASEAN4 and rest of South Asia lose, andthese losses are relatively small. As before the overall effects, as well as most of theindividual country effects, are dominated by the terms of trade changes. Two notableexceptions are China and MENA where the allocative efficiency effect dominates theterms of trade effect and changes the regional welfare outcome. In the case of China,this is due to a reduction in other processed food output, which shows a much higherrate of taxation than other sectors in this aggregation of the version 5 GTAP database.This gives rise to an efficiency loss. For MENA, the source of the large efficiency gainis due to the increase in imports. MENA’s imports of everything excepting programmecrops tend to increase only modestly. However, this region has very high rates ofprotection on many of these products imported from the EU and EFTA – indeed muchhigher than for most other products. Other processed food products is a case in point,with an average bilateral tariff of 165 per cent on imports from the EFTA region. Thuswhen other processed food products from EFTA increase, as a result of tradeliberalization in that region, there is a substantial efficiency gain for the MENA region.However, in the aggregate, these efficiency gains are only a small portion of the totaldeveloping country gains from the re-instrumentation experiment.
The breakout of the total regional terms of trade effects into their component parts in theremaining columns of Table 14 reveals that, unlike the domestic support scenario, theacross-the-board cut to market price support is most strongly influenced by the exportprice effect. With all OECD countries increasing their imports, and hence their exports,
29
the average price of OECD exports falls for most products. This depresses the worldaverage price of most products, leaving the developing countries with a favourableposition for their export prices, relative to the world average. Both the world price effectand the import price effect are still negative, but these are dominated by the strongpositive change in developing country export prices.
In order to explore the export price effect in greater detail, Table 15 presents thiscomponent of each country’s terms of trade at the individual commodity level. Here, wesee that, apart from the programme commodities, almost all the export price effects arepositive, reflecting the general tendency of OECD export prices to fall, relative to thoseof the developing countries. The total export price effect by commodity, summed overall the developing countries, shows the largest positive effects for other crops and otherprocessed food products. Table 15 also reports the total world and import price effects,by commodity, for the developing countries, as well as the total TOT effect (sum ofworld, export and import effects). On a commodity basis, the only negative entries inthis final column pertain to wheat and coarse grains. All other commodities show a totalTOT effect that is positive for the developing countries.
6 Summary and conclusions
Long term support for agricultural programme commodities in OECD countries,coupled with dis-protection in many developing countries, has left many of the latterincreasingly dependent on imports. In the historical overview section of this paper wereport trade specialization indexes over the past three decades for programme crops.These represent the grains and oilseeds which receive a large share of the domesticsupport in OECD countries. This measure is bounded between +1 and -1 and describesthe export (positive sign) and import (negative sign) orientation of each region. Withfew exceptions, these show substantial declines over this period. For example, Indonesiafalls from -0.57 to -0.88 and ASEAN4 falls from +0.58 to +0.20. Several regions showshifts from net exporter to net importer status. For example sub-Saharan Africa’s indexfalls from +0.39 in the 1965-75 period to -0.17 in the 1986-98 period, while the tradespecialization index for Latin America outside of Brazil, Argentina and Mexico fallsfrom 0.36 to -0.08. As these developing countries have come to rely on imports ofgrains and oilseeds from the subsidized OECD economies, they have become muchmore exposed to agricultural reforms that raise the prices of these specific products. Asa result, we find that an across-the-board, 50 per cent cut in all domestic support forOECD agriculture leads to welfare losses for most of the developing regions, as well asfor the combined total group of developing countries. The 50 per cent cut in domesticsupport also results in large declines in farm incomes in Europe, and, to a lesser degree,North America. This makes such a reform package an unlikely political event.
An alternative approach to reforming agricultural policies in the OECD would be tofocus on broad-based reductions in market price support. This has been occurring in anumber of OECD countries, most notably the EU where domestic support hasincreasingly replaced border measures. As demonstrated in this paper, the basiceconomic principles of agricultural support policies suggest that a shift from marketprice support to land-based payments could generate a ‘win-win’ outcome wherebyfarm incomes are maintained and world price distortions are reduced. This is thedirection charted by the OECD in its recent ‘Positive Reform Agenda’ for agriculture(OECD 2002). We formally examine such an agricultural reform scenario,
30
implementing a 50 per cent cut in market price support for OECD agriculture, with acompensating set of land payments designed to maintain farm income in each of themember economies. This comprehensive reform scenario results in increased welfarefor most developing countries, with gains on other commodities offsetting the terms oftrade losses from higher programme crop prices.
The preference for a continued focus on cuts in market price support, instead of shiftingthe emphasis to domestic support cuts is also reflected in two recent papers by otherauthors on this same general topic. Rae and Strutt (2002) conclude from their GTAP-based comparison between border measures and domestic support that improved marketaccess generates far greater trade and welfare gains than domestic support cuts. Thisleads them to propose that trade negotiators’ attention be focused squarely cuts toborder measures before turning any attention to domestic support.7 Hoekman et al.(2002) focus on developing country impacts of OECD agricultural policies using a verydifferent approach, but they reach the same conclusion as this paper.8 They find thatnamely that cuts to tariffs will generate much larger global welfare gains and positivegains to developing countries, whereas cuts to domestic support lead to smaller globalwelfare gains and losses for developing countries.
In summary, we conclude that developing countries will be well advised to focus theirefforts on improved market access to the OECD economies, while permitting thesewealthy economies to continue – indeed even increase – domestic support payments.Provided these increased domestic support payments are not linked to output or variableinputs, the trade-distorting effects are likely to be small, and they can be a rathereffective way of offsetting the potential losses that would otherwise be sustained byOECD farmers. This type of policy re-instrumentation will increase the probability thatsuch reforms will be deemed politically acceptable in the OECD member economies,while simultaneously increasing the likelihood that such reforms will also be beneficialto the developing economies.
References
Abler, D. G. and, J. S. Shortle (1992). ‘Environmental and Farm Commodity PolicyLinkages in the US and the EC’. European Review of Agricultural Economics, 19:197–217.
Anderson, K., B. Dimaranan, J. Francois, T. Hertel, B. Hoekman, and W. Martin(2001). ‘The Cost of Rich (and Poor) Country Protection to Developing Countries’.Journal of African Economies, 10 (3): 227–257.
Anderson, K., E. and M. D. Ingco (1999) ‘Integrating Agriculture into the WTO: TheNext Phase’. Paper prepared for the World Bank’s Conference on DevelopingCountries and the Millennium Round, Council Room, WTO Secretariat, CentreWilliam Rappard, Geneva (19-20 September).
7 Unlike this study, Rae and Strutt focus solely on cuts in domestic support provided through output andvariable input subsidies (their proxy for ‘amber box’ measures).
8 Their analysis is based on a highly disaggregate, econometric model that assumes products are perfectsubstitutes.
31
Dewbre, J., J. Anton, and W. Thompson (2001). ‘The Transfer Efficiency and TradeEffects of Direct Payments’. American Journal of Agricultural Economics, 83 (5):1204–1214.
Dimaranan, B. V. (2002). ‘Construction of the Protection Data Base’, in B. V. Dimaranan,and R. McDougall (2002), Global Trade, Assistance, and Production: The GTAP 5Data Base, Chapter 16.A. Center for Global Trade Analysis, Purdue University,West Lafayette, Indiana, USA.
Dimaranan, B. V., and R. McDougall (eds) (2002). Global Trade, Assistance, andProduction: The GTAP 5 Data Base. Center for Global Trade Analysis, PurdueUniversity, West Lafayette, Indiana, USA.
The Economist (2002). ‘Europe’s Farms: Will These Modest Proposals ProvokeMayhem Down on the Farm?’ (13-19 July): 42–43.
The Economist (2003). ‘The Case for the Defence’ (9 January).
Elbehri, A., M. D. Ingco, T. Hertel, and K. Pearson (1999). ‘Agriculture and WTO2000: Quantitative Assessment of Multilateral Liberalization of AgriculturalPolicies’. Paper presented at the Conference on Agriculture and the New TradeAgenda in the WTO 2000 Negotiations, 1-2 October, Geneva, Switzerland.
Floyd, J. E. (1965). ‘The Effects of Farm Price Supports on Returns to Land and Laborin Agriculture’. Journal of Political Economy, 73 (2): 148–158.
Francois, J., B. McDonald, and H. Nordstrom (1996). ‘The Uruguay Round: ANumerically-based Qualitative Assessment’, in W. Martin, and L. A. Winters, TheUruguay Round and the Developing Countries, Chapter 9. New York: CambridgeUniversity Press.
Frandsen, S., B. Gersfelt, and H. G. Jensen (2002). ‘Decoupling Support in Agriculture:Impacts of Redesigning European Agricultural Support’. Paper presented at the FifthAnnual Conference on Global Economic Analysis held in Taipei, Taiwan (5-7 June).
Frandsen, S., H. G. Jensen, and W. Yu (2001). ‘Domestic Support in AgricultureVersion 5 of the GTAP Database: The Case of the European Union’. Paper presentedat the Fourth Conference on Global Economic Analysis held at Purdue University,Indiana, USA (June).
Gehlhar, M. J., and F. J. Nelson (2001). ‘Treatment of Domestic Agricultural Supportand Implications for Reductions: The Case for the United States’. Paper presented atthe Fourth Conference on Global Economic Analysis held at Purdue University,Indiana, USA (June).
Gunter, L. F., Ki Hong Jeong, and F. C. White (1996). ‘Multiple Policy Goals in aTrade Model with Explicit Factor Markets’. American Journal of AgriculturalEconomics (May): 313–330.
Harrison, G. W., T. F. Rutherford, and D. G. Tarr (1996). ‘Quantifying the UruguayRound’, in W. Martin, and L. A. Winters, The Uruguay Round and the DevelopingCountries, Chapter 8. New York: Cambridge University Press.
Hertel, T. W. (1989). ‘Negotiating Reductions in Agricultural Support: Implications ofTechnology and Factor Mobility’. American Journal of Agricultural Economics(August): 559–573.
32
Hertel, T., and W. Martin (1999). ‘Would Developing Countries Gain from Inclusion ofManufactures in the WTO Negotiations?’ Paper prepared for the World Bank’sConference on Developing Countries and the Millennium Round, Council Room,WTO Secretariat, Centre William Rappard, Geneva (19-20 September).
Hertel, T., W. Martin, B. Dimaranan, and K. Yanagishima (1996). ‘Manufactures Tradein a Changing World Economy’, in W. Martin, and L. A. Winters, The UruguayRound and the Developing Countries, Chapter 7. New York: Cambridge UniversityPress.
Hoekman, B., F. Ng, and M. Olarreaga (2002). ‘Reducing Agricultural Tariffs versusDomestic Support: What’s More Important for Developing Countries?’. Paperpresented at the International Agricultural Trade Research Consortium Conferenceon ‘The Developing Countries, Agricultural Trade and the WTO’, Whistler (16-17June).
Huff, K. M., and T. W. Hertel (1996). ‘Decomposition of Welfare Effects in the GTAPModel’. GTAP Technical Paper No. 5, Center for Global Trade Analysis, PurdueUniversity, Indiana, USA.
Hummels, D. (1999). ‘Toward a Geography of Trade Costs’. Mimeo. University ofChicago, Illinois, USA.
Ianchovichina, E., and W. Martin (2002). ‘Economic Impacts of China’s Accession tothe WTO’. Mimeo. International Trade Department, World Bank.
Jensen, H. G. (2002). ‘Agricultural Domestic Support’, in B. V. Dimaranan, andR. McDougall, Global Trade, Assistance, and Production: The GTAP 5 Data Base,Chapter 16.E. Center for Global Trade Analysis, Purdue University, West Lafayette,Indiana, USA.
Martin, W., and L. A. Winters (eds) (1986). The Uruguay Round and the DevelopingEconomies. New York: Cambridge University Press.
McDougall, R. A. (1993). ‘Two Small Extensions to SALTER’. SALTER WorkingPaper No. 12, Industry Commission, Canberra, Australia.
Rae, A. N., and A. Strutt (2002). ‘The Current Round of Agricultural TradeNegotiations: Why Bother about Domestic Support?’ Paper presented at the FifthAnnual Conference on Global Economic Analysis held in Taipei, Taiwan (5-7 June).
Organization of Economic Cooperation and Development (OECD) (2001). MarketEffects of Crop Support Measures. Paris, France: OECD.
Organization of Economic Cooperation and Development (OECD) (2002). AgriculturalPolicies in OECD Countries: A Positive Agenda for Reform. Paris, France: OECD.
Thompson, W., and G. Smith (2002). The Medium-Term Impacts of TradeLiberalisation in OECD Countries on the Food Security of Non-Member Economies.OECD Directorate for Food, Agriculture and Fisheries, Paris, France.
Tyers, R., and K. Anderson (1992). Disarray in World Food Markets. CambridgeUniversity Press: Cambridge.
